Abstract
Stat electroencephalograms (sEEG) recorded over a one year period were analyzed to determine the rate of seizure or status epilepticus (SE) detection and the best predictors based on: ordering physician, clinical indication for study, and clinical history. All consecutive sEEG reports done over a year period at our institution were retrospectively reviewed. The following data were evaluated: sEEG findings, clinical history, clinical indication for study, requesting physician, location of patient, and demographics. Univariate analysis followed by a multivariate regression model analysis was performed. Of the 3,471 inpatient EEGs performed during the review period, 778 (22.4%) were sEEGs. 3.5% (n = 27) nonconvulsive status epilepticus (NCSE), 0.4% (n = 3) convulsive status epilepticus (CSE), and 1.1% (n = 9) had discrete electrographic seizures giving a total yield of 5.0% (39/778) patients with seizures or SE. A multivariate logistic retrospective model looking at ordering physician, clinical indication, and clinical history found that only clinical indications (overt continuous seizures/movements and witnessed seizure without return to baseline) were significant in the overall model. In our tertiary care institution sample, the rate of detecting status epilepticus or seizures among sEEG is low compared to prior studies. The best clinical predictors of finding SE or discrete seizures were overt continuous seizures/movements or witnessed seizure without return to baseline.